The impact of REACH on the environment and human health

Transcription

1 Report to DG Environment The impact of REACH on the environment and human health ENV.C.3/SER/2004/0042r Disclaimer: The contents of this report may not be regarded as stating an official position of the European Commission. The Commission accepts no responsibility or liability with regard to the use of the information in this report. Revised final report September 2005

2 The impact of REACH on the environment and human health ENV.C.3/SER/2004/0042r Agern Allé 5 DK-2970 Hørsholm Denmark Tel: Fax: fip@dhi.dk Web: Client DG Environment Client s representative Fabio Leone Project Authors The impact of REACH on the environment and human health Finn Pedersen Lise Samsøe-Petersen Kim Gustavson Lena Höglund (DTC) Phoebe Koundouri (University of Reading) David Pearce (University College London) Project No Date Approved by Torben Madsen 02 Revised final report KIG LSP TMA Final report FIP LSP TMA Revision Description By Checked Approved Date Key words Classification Open Internal Proprietary Distribution DG Environment DHI: FIP/LSP/KIG/ERA No. of copies 11 5

3 CONTENTS EXECUTIVE SUMMARY INTRODUCTION BACKGROUND OF THE STUDY Impact of the proposed REACH regulation Objectives of the study Anticipated functioning of REACH Effect of REACH on releases of chemicals IMPACT OF CHEMICALS ON ENVIRONMENT AND HUMAN HEALTH Retrospective analyses Predictive approaches METHODOLOGIES FOR EVALUATING ECONOMIC BENEFITS OF REACH A gallery of methodologies An ideal approach Willingness to pay and willingness to accept compensation Damage function approach based on past mistakes Avoided or saved costs approach Past studies Conclusions on current approaches SELECTED METHODOLOGIES The WTP approach The damage function approach based on past mistakes The avoided or saved costs approach Sewage treatment plants Drinking water purification Disposal of dredged sediment Disposal of sewage sludge on farmlands Cleaning of fish meal Estimating benefits of REACH General considerations Default benefit of REACH Overview of methodologies RESULTS Willingness to pay WTP for clean drinking water WTP for avoiding morbidity and mortality Damage function approach based on past mistakes ,2,4-trichlorobenzene in drinking water Nonylphenol in sewage sludge Tetrachloroethylene in ground water Polychlorinated biphenyls in fish Summary of benefits - damage function approach ERA/52856/Revised final report/ i

4 6.3 Avoided or saved costs approach Sewage treatment plants Drinking water purification Disposal of dredged sediment Sewage sludge incineration/disposal Cleaning of fish meal Summary of benefits - avoided or saved costs approach DISCUSSION AND CONCLUSIONS REFERENCES APPENDICES A B C D E COST-BENEFIT ANALYSIS OF REACH, SUMMARY OF STUDIES WTP AND BENEFITS TRANFER RANKING METHODS STUDIES ON CASE SUBSTANCES STUDIES ON RETROSPECTIVE ERRORS ERA/52856/Revised final report/ ii

5 EXECUTIVE SUMMARY A proposal for a new EU chemicals regulation on Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) was presented by the European Commission in October REACH will require that manufacturers and importers of chemicals register their chemicals, that registrations are evaluated by authorities, that certain substances of very high concern are authorised and that restrictions are imposed in cases where risks cannot be adequately controlled by other means. REACH will replace and consolidate into one single regulation large parts of the current chemicals legislation. The introduction of the new legislation will have an impact on human health and the environment and on society and business. Numerous studies have been conducted by the Commission, by national authorities and various stakeholders on the possible impact of REACH. Most of these studies have focused on the economic costs to industry, while only a few have dealt with the potential benefits. Therefore, the Commission has initiated and commissioned the current study on the impact of REACH on the environment and human health. The possibilities for estimating the potential benefits of REACH on the environment and on humans exposed via the environment suffer both from a lack of a sufficiently developed methodology and from a lack of data. In the present study, we have tried to circumvent these knowledge gaps by using three different methodologies for assessing potential benefits and to use a number of data at a screening level. Of course, this influences the reliability of the conclusions that can be drawn on the basis of the study. Three possible approaches have been identified that may be suitable for assessing the potential impact and benefits of REACH on the environment and humans exposed via the environment. These are: Willingness to pay (WTP) among the broad population for avoiding impacts of chemicals Damage function approach based on past mistakes where an empirical relationship between damage and cost might be established Avoided or saved costs approach where costs of mitigating current pollution is estimated as the upper limit for the possible benefit of REACH The WTP approach is seen as the economically correct way of estimating benefits. However, only two studies are available. One study from the UK elicits the population s willingness to pay for clean drinking water, while another study reviews the willingness to pay for avoiding health effects of chemicals pollution, in particular cancer. The information from the first study was used to estimate the potential benefits of REACH to 1,730 mill in year 2017 if only benefits to drinking water quality are considered. The study is not sufficient for extrapolating to the benefits of REACH on the environment in the whole of EU-25. It might be assumed that the population s WTP for environmental benefits is lower than for direct health benefits, while the WTP for avoiding serious health effects of chemicals pollution is much higher. Due to the very limited amount of input data, the results obtained are judged to be uncertain. ERA/52856/Revised final report/

6 The damage function approach based on past mistakes was tried out using four wellknown substances 1 as the basis and extrapolating to all other substances that may be affected by REACH. A system was established, which was intended to rank all substances based on their environmental and health properties in combination with tonnage; e.g. persistent toxic substances that are produced in large amounts are ranked very high. The ranking system was based on the EURAM method with input data obtained from the European Commission s IUCLID database and from the Danish EPA QSAR 2 database. The information from the IUCLID database is restricted to substances manufactured or imported in quantities above 10 tonnes/year and information on properties and amounts was provided for 8,031 substances. QSAR calculations were available for 45,452 discrete, organic substances among which 4,368 were recorded in IUCLID with information on quantities. All of the input data are uncertain and can only be used with caution. However, although the four substances selected as case substances are among those that are now restricted, the ranking showed that many other substances seem to be of similar or higher concern. Such a large number of substances cannot be assessed on a substance-by-substance approach, as any benefit resulting from reducing the release of one substance may be shadowed by impact from other substances. Instead, a conservative 10% reduction of costs due to REACH has been calculated. Due to the large uncertainty of the input data and the huge extrapolation, this approach is judged to be the weakest of the three approaches tried out in the current study. The avoided or saved costs approach was used to assess the current costs of mitigating the chemical pollution for a number of cases 3. The cost estimates for some of the cases are relatively robust, as it has been possible to obtain relatively detailed and precise information. This is in particular the case for purification of drinking water, disposal of dredged sediment and incineration of sewage sludge instead of disposing it on farmlands. Costs of building larger sewage treatment plants in order to obtain room for excess nitrification capacity due to toxic effects of chemicals in sewage water and costs of cleaning of fish products are considered weaker cases. From the cases, it is estimated that today the costs of measures already implemented for mitigating the impact of releases of chemicals are huge - in total up to 7 billion per year in 2005 for only those cases included in the study. Even assuming that the potential benefit of REACH would be only at 10%, the benefit is estimated to mill in year 2017, which over the next 25 years adds up to 2,800-9,000 mill. An overview of the results is given in Tables A-C below. Most of the estimates are based on an assumed efficiency of REACH in reducing general environmental contamination levels by 10%. 1 1,2,4-trichlorobenzene, nonylphenol, tetrachloroethylene and PCBs 2 Quantitative Structure-Activity Relationship (QSAR). QSAR are methods for estimating the toxicity and other properties of a chemical from its molecular structure. 3 Sewage treatment, drinking water purification, disposal of dredged sediment, sewage sludge incineration/disposal and cleaning of fish meal ERA/52856/Revised final report/

7 Table A Overview of potential benefits of REACH (values in mill ) determined as potentially saved costs (most robust approach) Case Building of sewage treatment plants Drinking water purification ,564 Disposal of dredged sediment ,450 (78-470)* (1, )* Sewage sludge 83 1,520 Cleaning of fish meal Total potential benefits for cases ,804-8,990 *) Based on 60% reduction of contaminated sediment. Table B Overview of potential benefits of REACH (values in mill ) determined as the population s willingness to pay (weaker approach) Case WTP for clean drinking water 1,730 34,000 Table C Overview of potential benefits of REACH (values in mill ) determined by extrapolation from case substances (weakest approach) Case Avoidance of severe health effects 210-2,500 4,000-50,000 Improved reuse of sewage sludge ,600 Total potential benefits for cases 226-2,633 4,300-52,600 It appears from the overview tables that the most robust approach results in the lowest estimate of benefits, while the weakest approach results in the largest estimate of benefits. However, the three different approaches estimate different costs and benefits, with the most robust approach estimating costs and benefits in relation to cleaning or handling of polluted matrices (water, sludge, sediment, fish products) and the weakest approach mainly estimating saved health costs. To obtain the best reflection of the different methodologies used and the level of uncertainty linked to the estimated impacts (i.e. indicative values), we preferred to keep them clearly separate. Thus, in conclusion, the potential benefit of REACH on the environment and humans exposed via the environment is estimated by use of a robust approach to as a minimum mill in year 2017 with a potential long-term benefit over the succeeding 25 years of 2,800-9,000 mill. These estimates are based on well-documented cases of costs in combination with assumptions on the potential benefits of REACH. Using much weaker approaches, the benefit arising from saved health costs is estimated to 200-2,500 mill in year 2017, which aggregated over 25 years corresponds to 4,000-50,000 mill. Once again, these values can only be seen as indicative values for the potential benefits of REACH on the environment and humans exposed via the environment. The values are based on a very weak data set; however, the best available. A ERA/52856/Revised final report/

8 more precise estimate would require generation of new data and a key role of REACH is to generate such data. We are particularly grateful for the input and comments received from two groups of experts, which were established for the purpose of reviewing this report. Their particular expertise in the domain of public health and environment, risk assessment, QSAR and environmental evaluation has been crucial to the successful completion of this study. ERA/52856/Revised final report/

9 1 INTRODUCTION The proposal for the REACH Regulation presented by the European Commission is currently being discussed in the Council of Ministers and in the European Parliament. It is the intention that REACH shall replace large parts of the current chemicals legislation. REACH will require that manufacturers and importers of chemicals register the chemicals, that registrations are evaluated by authorities, that certain substances of very high concern are authorised and that restrictions are imposed in cases where risks cannot be adequately controlled by other means. It is assumed that the REACH Regulation at the earliest will be adopted and enter into force during the year 2006 or more likely during Furthermore, an implementation period of 11 years is scheduled, which means that the full benefits of REACH will only be evident from the year 2017 at the earliest. 2 BACKGROUND OF THE STUDY 2.1 Impact of the proposed REACH regulation The Commission has prepared an assessment of the impact of REACH including a broad assessment of both the costs and the benefits (CEC 2003). This assessment was based on a number of analyses carried out by the Commission, by various contractors for the Commission and by various stakeholders including industry and NGOs. Most attention has been paid to the potential costs of implementing the REACH Regulation, while only a few studies have dealt with the potential benefits (cf. an overview of costbenefit analyses in Appendix A). Three studies on the benefits of REACH preceded the present study (RPA 2003, Postle et al. 2003, Pearce & Koundouri 2003). None of the studies evaluated environmental benefits, only benefits to health from either an economic or a risk management point of view. In summary, the assessments of the potential benefits of REACH are still uncertain and further work is needed in order to provide a more precise estimate of the benefits. This is in particular the case for benefits to the environment. 2.2 Objectives of the study The objectives of the study are: To assess the impact on the environment and humans exposed via the environment (i.e. excluding direct consumer exposure and occupational exposure) as a result of releases of chemicals; and To assess the possible long-term benefits of REACH in reducing such chemical threats. ERA/52856/Revised final report/

10 In the study appropriate methodologies were developed, available data were collected, and a number of case studies were conducted in order to create sufficient basic knowledge allowing the assessment of potential benefits of REACH. 2.3 Anticipated functioning of REACH It is a prerequisite for assessing the possible benefits of REACH that a sufficiently precise understanding of the functioning of REACH is established. For this particular study, the potential reduction of the releases of chemicals to the environment and the subsequent reduced exposure and effects on the environment and humans exposed via the environment are the focus. REACH may result in reduced releases to the environment through different instruments: Industry introduces additional Risk Management Measures (RMM) as a consequence of either having re-classified substances as a result of additional information on substance properties leading to additional S-phrases, or having identified risks by preparing a Chemical Safety Assessment (CSA) in relation to Registration of their chemicals. Use conditions are imposed as a result of an Authorisation obtained for certain uses of prioritised substances of very high concern. Restrictions on manufacturing, marketing or use as a result of the Restriction procedure. Of these instruments, the restriction procedure is essentially a continuation of the current restrictions directive (76/769/EEC). The influence of REACH on this work would relate to the speed of introducing new restrictions, but this is impossible to predict. Thus, the assumption is that REACH will have no or only minor influence on releases to the environment through this instrument. The authorisation procedure is new under REACH. Substances of very high concern will be prioritised for authorisation and specific conditions for granting an authorisation may be imposed. However, the identification of these substances is already well underway under the current legislation and the potential benefit of REACH would therefore pertain merely to additional substances that may be identified as a result of generation of new information. A general requirement for granting an authorisation is that risks are adequately controlled. In granting an authorisation, authorities will confirm that this is the case based on documentation submitted by the applicant. This means that, although the same requirement on adequate control of risks pertains to all other substances that are registered, a stricter control may apply to authorised substances than to other substances. On the other hand, for some substances where the socio-economic benefits exceed the risks, more lenient requirements may be allowed, and this can hardly be considered a benefit to the environment. Furthermore, it is anticipated that a maximum of substances can be handled through this instrument per year; in particular in the start-up phase. Thus, all in all it is difficult to assess whether the authorisation procedure will result in major benefits to the environment and, consequently, this is not included in the current study. ERA/52856/Revised final report/

11 Thus, the main impact of REACH on human health and the environment would probably arise as a result of the chemical safety assessment (CSA) conducted by manufacturers and importers prior to registration of their chemicals, and the subsequent implementation of necessary risk management measures by themselves and downstream users. A CSA is required as part of the registration dossier for substances manufactured or imported in a quantity of more than 10 tonnes per year per registrant. However, only for substances fulfilling the criteria for classification as dangerous or the PBT/vPvB 4 criteria, exposure assessment and risk characterisation will be part of the CSA. The exposure assessment and risk characterisation part of the CSA is the main vehicle for identifying a need for introducing additional risk management measures eventually leading to reduced releases. The chemicals registered under REACH are any new substances that have not previously been manufactured or imported in EU as well as the so-called phase-in substances (essentially what currently are named existing substances ). The requirements for new substances are, if anything, more lenient than today s requirements, which means that there will be no benefit of REACH for these substances. Thus, the main benefits of REACH can be assumed to be related to phase-in (existing) substances manufactured or imported in a quantity of more than 10 tonnes per year and meeting the criteria for classification as dangerous or the PBT/vPvB criteria. A number of substances within the scope of REACH are already covered by other legislation under which safe manufacture and use is or will be ensured: The existing substances review programme comprising 141 prioritised High Production Volume Chemicals eventually followed by development and implementation of a risk reduction strategy for substances for which risks are not sufficiently controlled PBT and vpvb substances identified among the existing HPVCs New notified substances Persistent Organic Pollutants (POP) (included in the Stockholm Convention) Ozone depleting substances included in the European list (EEC 2000b) and/or in the Montreal protocol (UNEP 2000) and sufficiently restricted Greenhouse gasses, including halocarbons (CFCs, HCFCsr, HFCs, PFCs and SF6) and sufficiently restricted Those substances or those uses of substances that are already comprised by current programmes are therefore excluded from the study. This includes the 141 substances and groups of substances covered by the current existing substances review programme, the PBT and vpvb substances identified among the HPVCs and currently undergoing evaluation, and substances the manufacturing, marketing and/or use of which are already restricted. However, it might also be anticipated that even without the introduction of REACH, a continued inclusion of more and more substances into the current review programme and eventual restrictions in the long run would result in the same level of environmental and human health protection as will be accelerated by REACH. Thus, the potential 4 PBT: Persistent, Bioaccumulative and Toxic; vpvb: very Persistent, very Bioaccumulative. ERA/52856/Revised final report/

12 benefit of REACH should be seen in relation to the level of protection that would have been reached within a certain time period. This can be illustrated as in Figure 2.1 below. Expected protection With REACH A B Without REACH 2017 Time T Figure 2.1 The potential benefit of REACH relative to expected protection without REACH The methodologies used in this report effectively measure area A + B in Figure 2.1. But the true benefits of REACH are given by area A only, i.e. the difference in risk and environmental quality with REACH as opposed to without REACH. Hence the methodologies could overstate the benefits of REACH. However, while a reasonable assumption for the year at which REACH has its full impact is 2017, there is no guidance at all on the year at which non-reach legislation will achieve the convergent level of risk reduction at time T in the diagram. Moreover, the slope of the non-reach line could be fairly flat taking the progress of the current existing substances programme into account, making the extent of overestimation small. The problem is that information simply does not exist on which to base any quantitative estimate. Finally, if there is overestimation of benefits, it will be offset by under-estimation due to the fact that the full range of health and ecosystem impacts are not captured by any of the methodologies considered. 2.4 Effect of REACH on releases of chemicals A number of routes of releases of chemicals to the environment from manufacture and use can be identified: Direct discharge of wastewater Discharge of wastewater via municipal sewage systems Ventilation outlet to atmosphere Evaporation to atmosphere Deposit of waste on unprotected soil Deposit of waste in municipal landfills Collection and treatment of hazardous waste ERA/52856/Revised final report/

13 Most of the release pathways are regulated more or less thoroughly already under today s legislation. Direct discharge of wastewater from larger industries is regulated by the Integrated Pollution Prevention and Control (IPPC) Directive 96/61/EC via discharge limits at EU level and/or at national level. However, this regulation deals mainly with the already known prioritised toxicants. Discharge from industries to municipal sewage treatment systems may be regulated at national level. No direct regulation of discharges from consumers is in place, although chemicals in consumer products are regulated by, e.g., Directive 76/769/EEC on restrictions of hazardous substances. Discharges from sewage treatment plants are often regulated at national level although seldom the releases of individual substances. Atmospheric releases of priority substances from larger industries are regulated by the IPPC Directive. The direct deposit of chemical waste on unprotected soil, which previously has caused large soil and groundwater pollution problems, has been reduced considerably in recent years by legislation and waste collection systems (e.g. 91/689/EEC on hazardous waste). Disposal of waste products, including sewage sludge, on farmlands is regulated with concentration limits for a few prioritised substances (e.g. 86/278/EEC on sewage sludge). The collection, handling, treatment and deposit of wastes are regulated at EU and national level by a number of EU Directives (e.g. 75/442/EEC on waste; 99/31/EC on deposit of waste (Landfill Directive); 2000/53/EC on End-of-Life Vehicles; and 2002/96/EC on Waste Electrical and Electronic Equipment ). Based on the considerations above, the starting assumption is that the potential benefit of REACH on the environment will mainly be a result of reduced releases of chemicals to water and to air. 3 IMPACT OF CHEMICALS ON ENVIRONMENT AND HUMAN HEALTH The release of chemicals to the environment may result in environmental effects on species in the environment and in health effects on humans exposed via the environment. The magnitude of such impacts is a function of the intrinsic hazards of the chemicals and the level of exposure of the environment and humans. The exposure levels are a result of the releases, which again are caused by the manufacture and use of the chemicals. This implies that there is a causal relationship between intrinsic hazards and manufacture and use of chemicals and the resulting impact of those chemicals. The resulting impact may be described in qualitative, quantitative and monetary terms. The cause-impact chain is illustrated in Figure 3.1. ERA/52856/Revised final report/

14 Chemical properties Production and use, amounts Releases Environmental fate, degradation & transport to receiving environment Impact on environment Impact on health from receiving environment Economic impact (damage costs) Figure 3.1 The cause-impact chain The impact of chemicals on the environment and human health via the environment is typically identified by use of approaches either based on case studies (which are by nature retrospective) or on risk characterisations where exposure is evaluated against noeffect-levels (developed to be predictive). Such approaches are described and discussed below, while approaches for monetising such impacts are described in the next chapter. 3.1 Retrospective analyses Most pollution effects of chemicals have been identified retrospectively on a case by case basis. Examples of cases of ecological impacts are numerous, e.g., contamination of groundwater and soil, decline in populations of birds of prey, reduced reproductive success of some species, and accumulated concentrations of pollutants in biota. There is also a large number of cases of effects on humans as a result of pollution, e.g. reduced fertility among the population in a specific area, increased incidence of foetal abnormalities and allergy incidents, as well as specific types of cancer. Some of these cases have been studied thoroughly and the causes are now well-known, while for others the cause-impact chain is still not fully understood. Examples of pollution cases where the causes of intoxications have been identified, and the cause-effect chain has been fully described, are plentiful and include the well-known effect of organochlorines on reproduction of top predators and the effects of mercury on the population in Minamata, Japan. Also the effects of CFC gases on the ozone layer are well known today. Common to these cases is that the investigation of the causes of these effects was initiated only when the effects became alarming due to their severity ERA/52856/Revised final report/

15 or extension. However, also common is the large time span between the identification of the effect and the full understanding of the causes. Similar examples are given in the EEA report on Late lessons learnt from early warnings (Harremöes et al. 2001). Even today new cases are found where it is not clear whether a cause-effect relationship exists resulting from manufacture and use of certain chemicals. For example, an increasing number of allergy cases appear among the European population, but the causes still remain to be identified. It is speculated that widely used chemicals may be at least part of the explanation. Another example is the relatively high concentrations of decabromodiphenyl ether that have recently been measured in samples of tissue and eggs of birds of prey from the UK, where neither the sources and environmental pathways nor the possible impact of the measured concentrations are currently known (Environment Agency 2004). A lesson learnt from such cases is that only when it is too late and pronounced effects become evident does the quest for the causes get initiated. And only when the causeeffect chain has been identified and documented are preventive measures identified and implemented (Harremöes et al. 2001). Thus, considering the long time span between the first releases and the implementation of corrective measures, a predictive approach has been introduced in the chemicals legislation. In conclusion, retrospective approaches are useful for elucidating the causal relationships between manufacture and use of chemicals and possible effects on the environment and humans. However, they require that the effects are so pronounced that they are identified as effects of chemical pollution and subsequently that the causes are identified by back-tracking. Luckily, corrective action has been taken for many of the pollution cases already under current legislation. 3.2 Predictive approaches The purpose of employing predictive approaches in the regulation of chemicals is to prevent the use of chemicals that leads to deleterious effects on the environment and human health. The main focus of these approaches has been to define at which exposure levels no unacceptable effects occur to the environment or human health and, based on that, to make sure that manufacture and use of chemicals do not lead to releases resulting in exceedance of these exposure levels. Thus, under the current chemicals legislation, the typical predictive approach for risk screening and assessment is based on the comparison of estimated exposure levels (concentrations, doses) with estimated no-effect-levels (e.g., Predicted No Effect Concentrations (PNEC) for environmental compartments, No Observed Adverse Effect Levels (NOAEL) for human populations). The Technical Guidance Document (TGD) on risk assessment (EC 2003) describes the elements of this approach in detail. The basic principles of this approach may be used at various levels of detail: Screening level. Approaches for screening of chemicals have been used for priority setting for regulatory purposes (e.g., the priority list of 141 substances for risk assessment under the Existing Substances Regulation (ESR) review programme) or ERA/52856/Revised final report/

16 for product assessment (e.g., which substances contribute significantly to the potential impact of a product in a lifecycle perspective). Generic risk assessment. Under the ESR programme, generic risk assessments are conducted at local and regional levels considering the manufacture, formulation, use (industrial/professional, private) and disposal stages of the lifecycle of chemicals. Site-specific risk assessment. Risk assessments of emissions from specific sites are conducted both under the ESR programme and at the local level in connection with issuing of production permits to industries, e.g. in accordance with the IPPC directive (Dir. 96/61/EC). Environment. A common feature of the approach is that it is based on an estimate of Predicted Environmental Concentrations (PEC) of chemicals. This is typically derived from information on produced or imported quantity and specific or generic information (as available) on use and release patterns and environmental fate and distribution. Often generic exposure models are used for estimating the PEC, but also more precise sitespecific models or even chemical measurements may be employed. The environmental concentrations are compared to a PNEC for the environmental compartment, which is a largely theoretical value defined as the concentration below which an unacceptable effect will most likely not occur (EC 2003). If the PEC/PNEC ratio (the risk quotient) is < 1, the risk of environmental effects is considered to be at an acceptable low level. However, although the PNEC is a limit above which unacceptable effects may occur, the PNEC does not give any indications of what types of ecological effects that may occur when this concentration is exceeded. Nor does it give information about the shape of the possible dose-response function for the ecosystem, i.e. how severe the effects may be in cases where PEC exceeds PNEC. No direct assessment of effects in air or the atmosphere is included in the current risk assessment approach. However, deposition of air pollutants to soil and water is considered, as well as the subsequent contribution to risks to these environments. Humans exposed via the environment. An indirect exposure of humans to contamination/pollution via environmental compartments may occur by inhalation of air, consumption of food (fish, meat, crops, (drinking) water), or dermal contact to soil or to polluted water during bathing or swimming. The common approach to perform exposure estimations is by estimating the total daily intake for humans. Normally, this is based on measurements or estimations of the PEC (for (surface) water, ground water, soil and air), the bioconcentration factor (measure for the bioaccumulation potential of a substance) and the biotransfer factor (chemical uptake by plants or animals) from the intake media, entailing the calculation of total daily intake (EC 2003). The calculated (daily) intake is then compared to a measure of effects, e.g. the No Observed Adverse Effects Level (NOAEL) or the Lowest Observed Adverse Effect Level (LOAEL). Critical effects for humans - acute and chronic - and their critical doses are identified or estimated, the critical effects being the adverse effects occurring at the lowest dose and the critical dose being the NOAEL. The NOAEL values may be established either directly from available experimental data or by applying one or more benchmark dose models. ERA/52856/Revised final report/

17 Methodological shortcomings Both screening level and generic risk assessment approaches comprise a number of assessment steps starting with produced amounts of chemicals to calculation of a (generic) risk quotient. As mentioned, the current methodological relationships between some of the steps in the cause-effect chain are relatively well established, while methodological approaches for other steps are lacking or, at best, only tentatively developed. The most important of these problems is that, very often, no relationship between exposure concentration and ecological or health effects (impact) has been established. As long as the exposure concentration is below the PNEC, it is assumed that the risk of ecotoxic effects is at an acceptable low level. In cases where the PEC exceeds the PNEC, a possibility of ecotoxic effects exists. However, the general problem with the PNEC approach is that PNEC is a no-effect-concentration at ecosystem level. Therefore, if the PNEC is exceeded there is no indication of what types of ecological effects that may occur (as illustrated in Figure 3.2). No information is conveyed about which species or functions of the ecosystem that may be affected or about the likelihood, possible magnitude or geographic scale of the occurrence of such effects. Ongoing developments within this field focus on concentration-effect relationships for individual key species in the ecosystem and the ecological impact of effects on populations of key species (e.g. Bradbury et al. 2004). Some knowledge is currently available, but not in a form that facilitates an ecological impact assessment exceeding individual case studies. Thus, the problems, which must be overcome in order to assess the environmental impact, include: The available information on production volume and use does not allow a localisation of estimated environmental concentrations and resulting effects. Our current risk assessment approach (i.e. the ratio between exposure concentration and no-effect-concentration) does not give any guidance on what type of environmental effects that may occur in case the exposure concentration exceeds the noeffect-concentration. ERA/52856/Revised final report/

18 Impact, effect on ecosystem / health Possible concentration-effect functions PNEC / NOAEL Concentration Figure 3.2 Illustration of the PNEC and types of dose-response functions, which are lacking for most chemicals With respect to health, a similar problem exists. In many cases we can define the equivalent to the PNEC (the NOAEL) but a dose-response relationship has not been defined to show how types of morbidity and premature mortality respond to the exceedance of the no-effect level. The problems which must be overcome in order to assess the human health impact include: Health damage will depend on localised exposure rather than general exposure. The numbers of people exposed at specific locations will generally not be known. However, assuming that an averaging approach can be used, average exposure levels could be estimated. These could then be compared with knowledge on probable effects at these exposure levels and the societal costs of identified effects. Of course human suffering cannot be determined individually but methods for quantifying societal costs have been developed. Nevertheless, in some cases, some relationship between measured concentrations and effects can be established from a hindsight perspective although the ecological or the human health impact may not have been predictable. Such examples can be used as a measure of the potential impact of chemicals with similar properties and emission, although the procedure is obviously subject to error. However, even if we could determine the types of effects that would occur and in which locations or environments, it may not be possible to monetise these impacts. ERA/52856/Revised final report/

19 4 METHODOLOGIES FOR EVALUATING ECONOMIC BENEFITS OF REACH 4.1 A gallery of methodologies Various methodologies for calculating economic costs of impact on environment and human health resulting from current manufacture and use of chemicals as well as methodologies for calculating economic benefits of REACH are outlined and discussed below An ideal approach The ideal measurement of economic impacts involves four end points : (a) health effects from exposure to ambient environmental concentrations (b) health effects from chemicals accumulated in food and water (c) health effects from exposure to chemicals in the workplace (d) loss of ecosystem functioning This could be termed the dose-response approach. For the present purpose, step (c) (workplace exposure) will not be considered. The practical implementation of the ideal approach requires that a number of essential procedures are used as follows: Estimate the chemical concentrations in each environmental compartment For assessing health effects Estimate dose-response relationships between those concentrations and human health premature mortality and morbidity Estimate the effects of REACH in terms of the likely reductions in environmental concentrations Estimate the change in the health effects arising Value those health effects using values of statistical life (VOSL), or value of life year (VOLY) and morbidity values For assessing environmental effects Estimate the change in concentrations as above Elicit people s willingness to pay for the improved environmental conditions as a result of implementing REACH For embodied health risks For risks embodied in ingested chemicals the procedures could be the same as above Difficulties in establishing relationships between ecological or health effects and economic costs The difficulties in securing economic damage measures are formidable, some of which were described above. Mainly they are: ERA/52856/Revised final report/

20 For the vast majority of the chemicals in question, no health dose-response information exists for the estimation of health effects. One approach is to take those relationships that are known and estimate the linkages for a few chemicals. Ecosystem effects are similarly not known with any precision. The effects of REACH on environmental concentrations are not known with any precision. Original willingness to pay (WTP) studies would ideally be required to secure valuations of ecosystem effects. However, the current study has neither the time nor the resources to conduct original WTP studies. Thus, it will be necessary to see standard values taken from other studies. This is unavoidably risky and the reliability of borrowing values ( benefits transfer ) has been seriously questioned in the economics literature (see, for example, Brouwer, 2000). The main problems lie not so much in the economic values themselves (the valuation of health effects is very well researched, ecosystem effects less so), but in the intermediary steps between estimating volume and toxicity of releases and the environmental and human health responses. Thus, in conclusion, current methodological approaches are not developed to a level where ecological or health effects of chemicals can be predicted. Neither are possibilities for monetising such impact sufficiently developed. However, a number of possible approaches exist or may be developed or adapted allowing a tentative assessment of economic impact and benefit of REACH Willingness to pay and willingness to accept compensation In all cases the theoretically correct economic value is either the population s willingness to pay (WTP) to avoid deleterious effects of chemicals, or the willingness to accept (WTA) compensation for tolerating the effects. The former applies when the population at large does not have the property rights to the future, improved state of the environment. The latter applies when the population does have the property right and hence compensation is required. The difference between WTP and WTA used not to be thought to be significant, but recent literature suggests that WTA may be several times (4-20 times) larger than WTP (Knetsch 1990, Hanemann 1991). This might be entirely rational in that, from a given starting point, consumers value losses more than gains. Moreover, consumers have poor understanding of the concept of WTA as they use it infrequently in their everyday life, whereas all consumers many times per day use the WTP concept when they purchase goods and services. For this reason it is considered safer (in terms of the accuracy of the results of a valuation study) to use WTP measures. Nevertheless, the property rights situation for chemicals is ambiguous. Directives such as the Water Framework Directive (Dir. 2000/60/EC) clearly give property rights (with caveats) to the population in respect of the future state of the environment (Pearce 2005). However, most economic analyses proceed with WTP rather than WTA because of the general presumption that individuals do not have property rights to a future, improved state of the environment. Usually, it is argued that property rights pertain to the existing level of environment: hence WTA would apply only if there are threats to make the existing environment worse through policy measures. Such situations are not uncommon - e.g. construction of a road or airport tends to infringe the rights to, say, the ERA/52856/Revised final report/

21 existing level of noise. However, in the REACH context, property rights are not so clear. It is possible to argue that the very existence of REACH implies a right to an environment with less chemical exposure. But in the absence of any clear guidance, and the fact that WTA estimates are harder to find than WTP estimates, we focus on WTP Damage function approach based on past mistakes The approach outlined here is close to the procedures used in some of the health economics, air pollution and risk analysis literature (Desvousges et al. 1998; Dolan et al. 2004) in which a damage function is established from the relationship between damage and the costs incurred by the damage. This approach has been used recently for monetising crime victims suffering, since harm done has been rated using qualityadjusted life years (QALYs, similar to DALYs), and the QALY-rated harm is then anchored on a form of harm where there are reasonably well established economic values (e.g. life lost) (Dolan et al. 2004). Similar approaches have been used for health impacts from other sources, again using QALYs or quality of wellbeing (QWB) scores (Desvousges et al. 1998). Employing the same approach for chemicals means that if the damage to the environment or human health caused by releases of one (or a few) chemicals as well as the economic costs of this damage is known, it might be possible to extrapolate to all other chemicals. This would require that a sufficiently solid relationship can be established between the damage and the costs. Assume that the damage described by an impact (Impact) on the environment or human health is established for a reasonably well-known chemical, for which some kind of economic analysis has either been done or could be done. Let the economic cost (Cost) from this reference chemical be X per unit of the chemical. Call this Cost R. Let the Cost for any chemical i be Cost i but for all chemicals bar the reference chemical we do not know the economic damages. So long as Impact is not an ordinal but a cardinal index 5 then we can compute the economic damage from any chemical as: Cost i Impact = Impact i R Cost R We then need some idea of the change in Impact arising from REACH. Call this ΔImpact. Then the benefit of REACH for any chemical is: Δ Cost i ΔImpact = Impact R i Cost R The obvious problems are: (a) We have no more information than the studies in the existing literature about the value for Impact. But we may have to accept that this is going to an expert judgement value. 5 An ordinal ranking simply says A is better than B which is better than C. A cardinal ranking tells how much better A is than B and how much B is better than C ERA/52856/Revised final report/

22 (b) If the economic valuation study for the reference chemical is not robust, all the chain-linked values will also be suspect. One way of addressing this problem is to see if we have another chemical with a reasonable valuation study, predict the damage cost of that chemical using the chain-linked approach, and then compare the predicted value with the actual value. (c) Problems caused by individual chemicals and the solution of such problems are not independent. This means that when a problem arising from one chemical is solved, at least part of similar problems caused by other chemicals may have been solved for the same cost. E.g., the cleaning of drinking water for one identified pollutant by use of active carbon will not only remove one chemical, but many different pollutants. Thus, there is a great risk in double-counting the benefits of REACH. (d) It is not clear that Impact will be a cardinal rather than an ordinal index. (e) We have no idea about the relationship between the Impact and the Cost other than Cost = f{impact}. The impact in itself is a function of manufacture, release, fate and effects, but whether a direct relationship between Impact and Cost can be established requires further research Avoided or saved costs approach Another approach to economic valuation is to use the costs of measures that have been introduced with the purpose of preventing, avoiding, repairing or mitigating damage caused by chemicals pollution. The starting point is that excess levels of chemicals in a specific environmental compartment may restrict the possibilities of using it, thereby implying a loss of potential future income or value and/or a cost for treatment or cleaning. Thus, if the soil in a garden is contaminated so that it cannot be used for playing by children or for growing vegetables, the value of the property decreases. This is a good way of measuring some of the damages - e.g. if REACH reduces the costs that water companies face in treating water, then their reduced costs are a legitimate way of measuring benefit of REACH. Contamination (caused by several chemicals) of specific environmental compartments or media, which reduces their usefulness for human purposes due to increased risks of unwanted effects, e.g. contamination of drinking water, may be reduced as a result of REACH. However, REACH may not affect all sources of contamination in these cases (e.g. nitrate or pesticides in drinking water), so the total costs of the problems will exceed the potential benefit of REACH. This approach could be termed the avoided or saved costs approach, as it describes the possible reduction in society s costs for mitigating chemicals pollution. However, it is important to differentiate this from the costs of restoring an ecosystem to some pre-damage state. This should not be regarded as the damage cost of the chemicals in question, because use of this procedure is not consistent with the principles of using WTP (WTA). In effect, we do not know if society is willing to incur the clean-up costs in question. Moreover, use of clean-up costs tends to make the benefit of clean-up identical with the cost of clean-up. This type of case includes contaminated surface and ground water, sewage sludge and contaminated soil. Examples of substances, which are contaminating environments but are regulated by other legislation than REACH, are pesticides and biocides as well as nitrates. ERA/52856/Revised final report/